Investigation of Polarized Light. 229 



if we do not with the first named investigators connect the 

 photo-electric process with the production of kathode-rays, it 

 remains to be explained why it is limited just to the kathode- 

 surface. To supply this deficiency a further assumption is neces- 

 sary which has recently been made by Prof. J. J. Thomson *, 

 namely, that the kathode-surface in contact with air, or what- 

 ever the gas may be, has a double electric charge, the positive 

 side formed by the molecules of the kathode, the negative by 

 molecules of the gas. 



If we now imagine that a ray of light so strikes the kathode- 

 surface of metal that the electric displacements in the ray have 

 a component at right angles to the surface, then electric 

 vibrations will be induced in the molecules of the metal in 

 which such a component will also be present. This will cause 

 the place of contact between the molecules of metal and gas, 

 so far as it belongs to the first, to take alternately positive and 

 negative charges in rapid succession. It is perhaps possible 

 that in that phase of the vibrations in which the electric 

 density is negative, the connexion of the metallic molecule 

 with the equally negative gas-molecule may be dissolved in 

 consequence of electrostatic repulsion, and the latter may be 

 driven off into the free gas-space, whilst another that now 

 takes its place communicates positive electricity to the 

 metallic molecule by contact with it, and itself takes a 

 negative charge. 



This view is supported by the circumstance that alkali- 

 metal cells, which, instead of rarefied hydrogen, contain the 

 much more strongly electro-negative gases oxygen or carbon 

 dioxide, are particularly sensitive to light, and that with 

 similar gaseous atmospheres the sensitiveness to light rises 

 with the electro-positive character of the kathode-metal. We 

 hope to return to these phenomena. 



It may certainly be urged against the view here expressed, 

 that the photo-electric action does not entirely cease when the 

 electric displacements take place parallel to the kathode- 

 surface, and when consequently the normal component is zero. 

 It appears to us to be of great importance to our knowledge 

 of the photo-electric process to make further investigations 

 into its dependence upon the angle of incidence, using, of 

 course, polarized light. 



* J. J. Thomson, Phil. Mag. xxxvii. p. 356 (1894). 



